COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\\M0BASL30\M1BASI30\.
C00008 ENDMK
C⊗;
\\M0BASL30;\M1BASI30;\.
\F0\;
\J	The subcommittee finds itself unable to decide conclusively,
from the evidence available, either what is the best computer to afford Stanford
the advantages of interactive computing or what is the best way
of operating the computer.  Each of the three main schemes has its
advocates, and the advantages of each scheme depend partly on
issues which cannot be evaluated experimentally at this time.
Some of the issues which remain unresolved are the following:

	1. What level of staffing is required to make the computers
available to different classes of users?  In general, the PDP-10
advocates propose minimal staffing relying on the time-sharing systems
themselves and their documentation.  SCIP, on the other hand, considers
that a high level of user services is required to make the system
available to a large population of users.  Some say that minimal
staffing would benefit research users and harm students, while the
Computer Science Department representatives say that teaching
assistants can adequately advise the students.  The staffing issue
is very important, because staff costs are about half of SCIP's
present costs and would be an even greater fraction of costs if a
fully amortized 360/67 is kept or if a PDP-10 is used and supported
with a staff of SCIP's present size.

	2. Which users would prefer the services offered by a PDP-10
system and which would prefer the ORVYL services?  How would prices
influence their preference?

	3. At what cost can the services be offered?

	4. Are there still other computer systems that would be
attractive to substantial groups of Stanford users?

	In view of these uncertainties, the subcommittee makes the
following organizational recommendations:

	1. The Provost computer funds become convertible and
usable on any machine including non-Stanford machines for which
a reasonable case can be made.  These funds amount to $700,000
per year and their commitment to the amortization of the purchase
of the 360/67 should be cnsidered terminated after April 1974.
The 370/158 was not bought to serve this community, and this
community should not be obligated to support it.

	2. In order to facilitate use of these funds, Associate
Provost for Computing should negotiate open-ended arrangements
for general use of a PDP-10 computer with TENEX.  Possibilities
include the IMSSS machine and machines operated by TYMSHARE.
The computer time should be purchased on a no-consulting-services
basis, and Stanford staffing should be held to a level that will
allow a markup of the price paid to the vendor of no more than
15%.  A user requiring additional assistance will be able to find
consultants for a price.  

	3. It is recommended that SCIP \F1unbundle\F0 its user services.
Specifically, it should charge for its courses and for consulting.
It may offer courses and consulting for users of the outside machines,
but this help should be charged for.

	4. The formation of users' groups for the outside machine
services is encouraged.

	\F1To recapitulate\F0: A free market for a few years in computer
services will tell us where the demand really lies and what services
are really valued by the users.  Otherwise, we shall have to haggle
over every job and service and try to determine what group of users
is most benefited by the service.\.